lnd.xprv/networktest.go

895 lines
25 KiB
Go
Raw Normal View History

package main
import (
"encoding/hex"
"fmt"
"io/ioutil"
"log"
"net"
"os"
"os/exec"
"path/filepath"
"runtime"
"strconv"
"sync"
"time"
"golang.org/x/net/context"
"google.golang.org/grpc"
"google.golang.org/grpc/grpclog"
"bytes"
"github.com/go-errors/errors"
"github.com/lightningnetwork/lnd/lnrpc"
"github.com/roasbeef/btcd/chaincfg"
"github.com/roasbeef/btcd/rpctest"
"github.com/roasbeef/btcd/txscript"
"github.com/roasbeef/btcd/wire"
"github.com/roasbeef/btcrpcclient"
"github.com/roasbeef/btcutil"
)
var (
// numActiveNodes is the number of active nodes within the test network.
numActiveNodes = 0
// defaultNodePort is the initial p2p port which will be used by the
// first created lightning node to listen on for incoming p2p
// connections. Subsequent allocated ports for future lighting nodes
// instances will be monotonically increasing odd numbers calculated as
// such: defaultP2pPort + (2 * harness.nodeNum).
defaultNodePort = 19555
// defaultClientPort is the initial rpc port which will be used by the
// first created lightning node to listen on for incoming rpc
// connections. Subsequent allocated ports for future rpc harness
// instances will be monotonically increasing even numbers calculated
// as such: defaultP2pPort + (2 * harness.nodeNum).
defaultClientPort = 19556
harnessNetParams = &chaincfg.SimNetParams
)
// generateListeningPorts returns two strings representing ports to listen on
// designated for the current lightning network test. If there haven't been any
// test instances created, the default ports are used. Otherwise, in order to
// support multiple test nodes running at once, the p2p and rpc port are
// incremented after each initialization.
func generateListeningPorts() (int, int) {
var p2p, rpc int
if numActiveNodes == 0 {
p2p = defaultNodePort
rpc = defaultClientPort
} else {
p2p = defaultNodePort + (2 * numActiveNodes)
rpc = defaultClientPort + (2 * numActiveNodes)
}
return p2p, rpc
}
// lightningNode represents an instance of lnd running within our test network
// harness. Each lightningNode instance also fully embedds an RPC client in
2016-10-15 16:18:38 +03:00
// order to pragmatically drive the node.
type lightningNode struct {
cfg *config
rpcAddr string
p2pAddr string
rpcCert []byte
nodeId int
// LightningId is the ID, or the sha256 of the node's identity public
// key. This field will only be populated once the node itself has been
// started via the start() method.
LightningID [32]byte
// PubKey is the serialized compressed identity public key of the node.
// This field will only be populated once the node itself has been
// started via the start() method.
PubKey [33]byte
PubKeyStr string
cmd *exec.Cmd
pidFile string
extraArgs []string
lnrpc.LightningClient
}
// newLightningNode creates a new test lightning node instance from the passed
// rpc config and slice of extra arguments.
func newLightningNode(rpcConfig *btcrpcclient.ConnConfig, lndArgs []string) (*lightningNode, error) {
var err error
cfg := &config{
RPCHost: "127.0.0.1",
RPCUser: rpcConfig.User,
RPCPass: rpcConfig.Pass,
}
nodeNum := numActiveNodes
cfg.DataDir, err = ioutil.TempDir("", "lndtest-data")
if err != nil {
return nil, err
}
cfg.LogDir, err = ioutil.TempDir("", "lndtest-log")
if err != nil {
return nil, err
}
cfg.PeerPort, cfg.RPCPort = generateListeningPorts()
numActiveNodes++
return &lightningNode{
cfg: cfg,
p2pAddr: net.JoinHostPort("127.0.0.1", strconv.Itoa(cfg.PeerPort)),
rpcAddr: net.JoinHostPort("127.0.0.1", strconv.Itoa(cfg.RPCPort)),
rpcCert: rpcConfig.Certificates,
nodeId: nodeNum,
extraArgs: lndArgs,
}, nil
}
// genArgs generates a slice of command line arguments from the lightningNode's
// current config struct.
func (l *lightningNode) genArgs() []string {
var args []string
encodedCert := hex.EncodeToString(l.rpcCert)
args = append(args, fmt.Sprintf("--btcdhost=%v", l.cfg.RPCHost))
args = append(args, fmt.Sprintf("--rpcuser=%v", l.cfg.RPCUser))
args = append(args, fmt.Sprintf("--rpcpass=%v", l.cfg.RPCPass))
args = append(args, fmt.Sprintf("--rawrpccert=%v", encodedCert))
args = append(args, fmt.Sprintf("--rpcport=%v", l.cfg.RPCPort))
args = append(args, fmt.Sprintf("--peerport=%v", l.cfg.PeerPort))
args = append(args, fmt.Sprintf("--logdir=%v", l.cfg.LogDir))
args = append(args, fmt.Sprintf("--datadir=%v", l.cfg.DataDir))
args = append(args, fmt.Sprintf("--simnet"))
if l.extraArgs != nil {
args = append(args, l.extraArgs...)
}
return args
}
// start launches a new process running lnd. Additionally, the PID of the
// launched process is saved in order to possibly kill the process forcibly
// later.
func (l *lightningNode) start(lndError chan error) error {
args := l.genArgs()
l.cmd = exec.Command("lnd", args...)
// Redirect stderr output to buffer
var errb bytes.Buffer
l.cmd.Stderr = &errb
if err := l.cmd.Start(); err != nil {
return err
}
// Launch a new goroutine which that bubbles up any potential fatal
// process errors to the goroutine running the tests.
go func() {
if err := l.cmd.Wait(); err != nil {
lndError <- errors.New(errb.String())
}
}()
pid, err := os.Create(filepath.Join(l.cfg.DataDir,
fmt.Sprintf("%s.pid", l.nodeId)))
if err != nil {
return err
}
l.pidFile = pid.Name()
if _, err = fmt.Fprintf(pid, "%s\n", l.cmd.Process.Pid); err != nil {
return err
}
if err := pid.Close(); err != nil {
return err
}
opts := []grpc.DialOption{
grpc.WithInsecure(),
grpc.WithBlock(),
grpc.WithTimeout(time.Second * 20),
}
conn, err := grpc.Dial(l.rpcAddr, opts...)
if err != nil {
return nil
}
l.LightningClient = lnrpc.NewLightningClient(conn)
// Obtain the lnid of this node for quick identification purposes.
ctxb := context.Background()
info, err := l.GetInfo(ctxb, &lnrpc.GetInfoRequest{})
if err != nil {
return nil
}
lnID, err := hex.DecodeString(info.LightningId)
if err != nil {
return err
}
copy(l.LightningID[:], lnID)
l.PubKeyStr = info.IdentityPubkey
pubkey, err := hex.DecodeString(info.IdentityPubkey)
if err != nil {
return err
}
copy(l.PubKey[:], pubkey)
return nil
}
// cleanup cleans up all the temporary files created by the node's process.
func (l *lightningNode) cleanup() error {
dirs := []string{
l.cfg.LogDir,
l.cfg.DataDir,
}
var err error
for _, dir := range dirs {
if err = os.RemoveAll(dir); err != nil {
log.Printf("Cannot remove dir %s: %v", dir, err)
}
}
return err
}
// stop attempts to stop the active lnd process.
func (l *lightningNode) stop() error {
// We should skip node stop in case:
// - start of the node wasn't initiated
// - process wasn't spawned
// - process already finished
processFinished := l.cmd.ProcessState != nil &&
l.cmd.ProcessState.Exited()
if l.cmd == nil || l.cmd.Process == nil || processFinished {
return nil
}
if runtime.GOOS == "windows" {
return l.cmd.Process.Signal(os.Kill)
}
return l.cmd.Process.Signal(os.Interrupt)
}
// shutdown stops the active lnd process and clean up any temporary directories
// created along the way.
func (l *lightningNode) shutdown() error {
if err := l.stop(); err != nil {
return err
}
if err := l.cleanup(); err != nil {
return err
}
return nil
}
// networkHarness is an integration testing harness for the lightning network.
// The harness by default is created with two active nodes on the network:
// Alice and Bob.
type networkHarness struct {
rpcConfig btcrpcclient.ConnConfig
netParams *chaincfg.Params
Miner *rpctest.Harness
activeNodes map[int]*lightningNode
// Alice and Bob are the initial seeder nodes that are automatically
// created to be the initial participants of the test network.
Alice *lightningNode
Bob *lightningNode
seenTxns chan wire.ShaHash
watchRequests chan *watchRequest
// Channel for transmitting stderr output from failed lightning node
// to main process.
lndErrorChan chan error
sync.Mutex
}
// newNetworkHarness creates a new network test harness.
// TODO(roasbeef): add option to use golang's build library to a binary of the
// current repo. This'll save developers from having to manually `go install`
// within the repo each time before changes
func newNetworkHarness() (*networkHarness, error) {
return &networkHarness{
activeNodes: make(map[int]*lightningNode),
seenTxns: make(chan wire.ShaHash),
watchRequests: make(chan *watchRequest),
lndErrorChan: make(chan error),
}, nil
}
// InitializeSeedNodes initialized alice and bob nodes given an already
// running instance of btcd's rpctest harness and extra command line flags,
// which should be formatted properly - "--arg=value".
func (n *networkHarness) InitializeSeedNodes(r *rpctest.Harness, lndArgs []string) error {
nodeConfig := r.RPCConfig()
n.netParams = r.ActiveNet
n.Miner = r
n.rpcConfig = nodeConfig
var err error
n.Alice, err = newLightningNode(&nodeConfig, lndArgs)
if err != nil {
return err
}
n.Bob, err = newLightningNode(&nodeConfig, lndArgs)
if err != nil {
return err
}
n.activeNodes[n.Alice.nodeId] = n.Alice
n.activeNodes[n.Bob.nodeId] = n.Bob
return err
}
// ProcessErrors returns a channel used for reporting any fatal process errors.
// If any of the active nodes within the harness' test network incur a fatal
// error, that error is sent over this channel.
func (n *networkHarness) ProcessErrors() chan error {
return n.lndErrorChan
}
// fakeLogger is a fake grpclog.Logger implementation. This is used to stop
// grpc's logger from printing directly to stdout.
type fakeLogger struct{}
func (f *fakeLogger) Fatal(args ...interface{}) {}
func (f *fakeLogger) Fatalf(format string, args ...interface{}) {}
func (f *fakeLogger) Fatalln(args ...interface{}) {}
func (f *fakeLogger) Print(args ...interface{}) {}
func (f *fakeLogger) Printf(format string, args ...interface{}) {}
func (f *fakeLogger) Println(args ...interface{}) {}
// SetUp starts the initial seeder nodes within the test harness. The initial
// node's wallets will be funded wallets with ten 1 BTC outputs each. Finally
// rpc clients capable of communicating with the initial seeder nodes are
// created.
func (n *networkHarness) SetUp() error {
// Swap out grpc's default logger with out fake logger which drops the
// statements on the floor.
grpclog.SetLogger(&fakeLogger{})
// Start the initial seeder nodes within the test network, then connect
// their respective RPC clients.
var wg sync.WaitGroup
errChan := make(chan error, 2)
wg.Add(2)
go func() {
var err error
defer wg.Done()
if err = n.Alice.start(n.lndErrorChan); err != nil {
errChan <- err
return
}
}()
go func() {
var err error
defer wg.Done()
if err = n.Bob.start(n.lndErrorChan); err != nil {
errChan <- err
return
}
}()
wg.Wait()
select {
case err := <-errChan:
return err
default:
}
// Load up the wallets of the seeder nodes with 10 outputs of 1 BTC
// each.
ctxb := context.Background()
addrReq := &lnrpc.NewAddressRequest{lnrpc.NewAddressRequest_WITNESS_PUBKEY_HASH}
clients := []lnrpc.LightningClient{n.Alice, n.Bob}
for _, client := range clients {
for i := 0; i < 10; i++ {
resp, err := client.NewAddress(ctxb, addrReq)
if err != nil {
return err
}
addr, err := btcutil.DecodeAddress(resp.Address, n.netParams)
if err != nil {
return err
}
addrScript, err := txscript.PayToAddrScript(addr)
if err != nil {
return err
}
output := &wire.TxOut{
PkScript: addrScript,
Value: btcutil.SatoshiPerBitcoin,
}
if _, err := n.Miner.CoinbaseSpend([]*wire.TxOut{output}); err != nil {
return err
}
}
}
// We generate several blocks in order to give the outputs created
// above a good number of confirmations.
if _, err := n.Miner.Node.Generate(10); err != nil {
return err
}
// Finally, make a connection between both of the nodes.
if err := n.ConnectNodes(ctxb, n.Alice, n.Bob); err != nil {
return err
}
// Now block until both wallets have fully synced up.
expectedBalance := btcutil.Amount(btcutil.SatoshiPerBitcoin * 10).ToBTC()
balReq := &lnrpc.WalletBalanceRequest{}
balanceTicker := time.Tick(time.Millisecond * 50)
out:
for {
select {
case <-balanceTicker:
aliceResp, err := n.Alice.WalletBalance(ctxb, balReq)
if err != nil {
return err
}
bobResp, err := n.Bob.WalletBalance(ctxb, balReq)
if err != nil {
return err
}
if aliceResp.Balance == expectedBalance &&
bobResp.Balance == expectedBalance {
break out
}
case <-time.After(time.Second * 30):
return fmt.Errorf("balances not synced after deadline")
}
}
// Now that the initial test network has been initialized, launch the
// network wather.
go n.networkWatcher()
return nil
}
// TearDownAll tears down all active nodes within the test lightning network.
func (n *networkHarness) TearDownAll() error {
for _, node := range n.activeNodes {
if err := node.shutdown(); err != nil {
return err
}
}
return nil
}
// NewNode fully initializes a returns a new lightningNode binded to the
// current instance of the network harness. The created node is running, but
// not yet connected to other nodes within the network.
func (n *networkHarness) NewNode(extraArgs []string) (*lightningNode, error) {
n.Lock()
defer n.Unlock()
node, err := newLightningNode(&n.rpcConfig, extraArgs)
if err != nil {
return nil, err
}
if err := node.start(n.lndErrorChan); err != nil {
return nil, err
}
n.activeNodes[node.nodeId] = node
return node, nil
}
// ConnectNodes establishes an encrypted+authenticated p2p connection from node
// a towards node b.
func (n *networkHarness) ConnectNodes(ctx context.Context, a, b *lightningNode) error {
bobInfo, err := b.GetInfo(ctx, &lnrpc.GetInfoRequest{})
if err != nil {
return err
}
req := &lnrpc.ConnectPeerRequest{
Addr: &lnrpc.LightningAddress{
PubKeyHash: bobInfo.IdentityAddress,
Host: b.p2pAddr,
},
}
if _, err := a.ConnectPeer(ctx, req); err != nil {
return err
}
return nil
}
// TODO(roasbeef): add a WithChannel higher-order function?
// * python-like context manager w.r.t using a channel within a test
// * possibly adds more funds to the target wallet if the funds are not
// enough
// watchRequest encapsulates a request to the harness' network watcher to
// dispatch a notification once a transaction with the target txid is seen
// within the test network.
type watchRequest struct {
txid wire.ShaHash
eventChan chan struct{}
}
// networkWatcher is a goroutine which accepts async notification requests for
// the broadcast of a target transaction, and then dispatches the transaction
// once its seen on the network.
func (n *networkHarness) networkWatcher() {
seenTxns := make(map[wire.ShaHash]struct{})
clients := make(map[wire.ShaHash][]chan struct{})
for {
select {
case req := <-n.watchRequests:
// If we've already seen this transaction, then
// immediately dispatch the request. Otherwise, append
// to the list of clients who are watching for the
// broadcast of this transaction.
if _, ok := seenTxns[req.txid]; ok {
close(req.eventChan)
} else {
clients[req.txid] = append(clients[req.txid], req.eventChan)
}
case txid := <-n.seenTxns:
// Add this txid to our set of "seen" transactions. So
// we're able to dispatch any notifications for this
// txid which arrive *after* it's seen within the
// network.
seenTxns[txid] = struct{}{}
// If there isn't a registered notification for this
// transaction then ignore it.
txClients, ok := clients[txid]
if !ok {
continue
}
// Otherwise, dispatch the notification to all clients,
// cleaning up the now un-needed state.
for _, client := range txClients {
close(client)
}
delete(clients, txid)
}
}
}
// OnTxAccepted is a callback to be called each time a new transaction has been
// broadcast on the network.
func (n *networkHarness) OnTxAccepted(hash *wire.ShaHash, amt btcutil.Amount) {
go func() {
n.seenTxns <- *hash
}()
}
// WaitForTxBroadcast blocks until the target txid is seen on the network. If
// the transaction isn't seen within the network before the passed timeout,
// then an error is returend.
func (n *networkHarness) WaitForTxBroadcast(ctx context.Context, txid wire.ShaHash) error {
eventChan := make(chan struct{})
n.watchRequests <- &watchRequest{txid, eventChan}
select {
case <-eventChan:
return nil
case <-ctx.Done():
return fmt.Errorf("tx not seen before context timeout")
}
}
2016-10-15 16:18:38 +03:00
// OpenChannel attempts to open a channel between srcNode and destNode with the
// passed channel funding parameters. If the passed context has a timeout, then
2016-10-15 16:18:38 +03:00
// if the timeout is reached before the channel pending notification is
// received, an error is returned.
func (n *networkHarness) OpenChannel(ctx context.Context,
srcNode, destNode *lightningNode, amt btcutil.Amount,
numConfs uint32) (lnrpc.Lightning_OpenChannelClient, error) {
openReq := &lnrpc.OpenChannelRequest{
TargetNode: destNode.LightningID[:],
LocalFundingAmount: int64(amt),
NumConfs: numConfs,
}
respStream, err := srcNode.OpenChannel(ctx, openReq)
if err != nil {
return nil, fmt.Errorf("unable to open channel between "+
"alice and bob: %v", err)
}
chanOpen := make(chan struct{})
errChan := make(chan error)
go func() {
// Consume the "channel pending" update. This waits until the node
// notifies us that the final message in the channel funding workflow
// has been sent to the remote node.
resp, err := respStream.Recv()
if err != nil {
errChan <- err
return
}
if _, ok := resp.Update.(*lnrpc.OpenStatusUpdate_ChanPending); !ok {
errChan <- fmt.Errorf("expected channel pending update, "+
"instead got %v", resp)
return
}
close(chanOpen)
}()
select {
case <-ctx.Done():
return nil, fmt.Errorf("timeout reached before chan pending " +
"update sent")
case err := <-errChan:
return nil, err
case <-chanOpen:
return respStream, nil
}
}
// WaitForChannelOpen waits for a notification that a channel is open by
// consuming a message from the past open channel stream. If the passed context
// has a timeout, then if the timeout is reached before the channel has been
// opened, then an error is returned.
func (n *networkHarness) WaitForChannelOpen(ctx context.Context,
openChanStream lnrpc.Lightning_OpenChannelClient) (*lnrpc.ChannelPoint, error) {
errChan := make(chan error)
respChan := make(chan *lnrpc.ChannelPoint)
go func() {
resp, err := openChanStream.Recv()
if err != nil {
errChan <- fmt.Errorf("unable to read rpc resp: %v", err)
return
}
fundingResp, ok := resp.Update.(*lnrpc.OpenStatusUpdate_ChanOpen)
if !ok {
errChan <- fmt.Errorf("expected channel open update, "+
"instead got %v", resp)
return
}
respChan <- fundingResp.ChanOpen.ChannelPoint
}()
select {
case <-ctx.Done():
return nil, fmt.Errorf("timeout reached while waiting for " +
"channel open")
case err := <-errChan:
return nil, err
case chanPoint := <-respChan:
return chanPoint, nil
}
}
// CloseChannel close channel attempts to close the channel indicated by the
// passed channel point, initiated by the passed lnNode. If the passed context
// has a timeout, then if the timeout is reached before the channel close is
// pending, then an error is returned.
func (n *networkHarness) CloseChannel(ctx context.Context,
lnNode *lightningNode, cp *lnrpc.ChannelPoint,
force bool) (lnrpc.Lightning_CloseChannelClient, error) {
closeReq := &lnrpc.CloseChannelRequest{
ChannelPoint: cp,
Force: force,
}
closeRespStream, err := lnNode.CloseChannel(ctx, closeReq)
if err != nil {
return nil, fmt.Errorf("unable to close channel: %v", err)
}
errChan := make(chan error)
fin := make(chan struct{})
go func() {
// Consume the "channel close" update in order to wait for the closing
// transaction to be broadcast, then wait for the closing tx to be seen
// within the network.
closeResp, err := closeRespStream.Recv()
if err != nil {
errChan <- err
return
}
pendingClose, ok := closeResp.Update.(*lnrpc.CloseStatusUpdate_ClosePending)
if !ok {
errChan <- fmt.Errorf("expected channel close update, "+
"instead got %v", pendingClose)
return
}
closeTxid, err := wire.NewShaHash(pendingClose.ClosePending.Txid)
if err != nil {
errChan <- err
return
}
if err := n.WaitForTxBroadcast(ctx, *closeTxid); err != nil {
errChan <- err
return
}
close(fin)
}()
// Wait until either the deadline for the context expires, an error
// occurs, or the channel close update is received.
select {
case <-ctx.Done():
return nil, fmt.Errorf("timeout reached before channel close " +
"initiated")
case err := <-errChan:
return nil, err
case <-fin:
return closeRespStream, nil
}
}
// WaitForChannelClose waits for a notification from the passed channel close
// stream that the node has deemed the channel has been fully closed. If the
// passed context has a timeout, then if the timeout is reached before the
// notification is received then an error is returned.
func (n *networkHarness) WaitForChannelClose(ctx context.Context,
closeChanStream lnrpc.Lightning_CloseChannelClient) (*wire.ShaHash, error) {
errChan := make(chan error)
updateChan := make(chan *lnrpc.CloseStatusUpdate_ChanClose)
go func() {
closeResp, err := closeChanStream.Recv()
if err != nil {
errChan <- err
return
}
closeFin, ok := closeResp.Update.(*lnrpc.CloseStatusUpdate_ChanClose)
if !ok {
errChan <- fmt.Errorf("expected channel close update, "+
"instead got %v", closeFin)
return
}
updateChan <- closeFin
}()
// Wait until either the deadline for the context expires, an error
// occurs, or the channel close update is received.
select {
case <-ctx.Done():
return nil, fmt.Errorf("timeout reached before update sent")
case err := <-errChan:
return nil, err
case update := <-updateChan:
return wire.NewShaHash(update.ChanClose.ClosingTxid)
}
}
// AssertChannelExists asserts that an active channel identified by
// channelPoint is known to exist from the point-of-view of node..
func (n *networkHarness) AssertChannelExists(ctx context.Context,
node *lightningNode, chanPoint *wire.OutPoint) error {
req := &lnrpc.ListChannelsRequest{}
resp, err := node.ListChannels(ctx, req)
if err != nil {
return fmt.Errorf("unable fetch node's channels: %v", err)
}
for _, channel := range resp.Channels {
if channel.ChannelPoint == chanPoint.String() {
return nil
}
}
return fmt.Errorf("channel not found")
}
// DumpLogs reads the current logs generated by the passed node, and returns
// the logs as a single string. This function is useful for examining the logs
// of a particular node in the case of a test failure.
// Logs from lightning node being generated with delay - you should
// add time.Sleep() in order to get all logs.
func (n *networkHarness) DumpLogs(node *lightningNode) (string, error) {
logFile := fmt.Sprintf("%v/simnet/lnd.log", node.cfg.LogDir)
buf, err := ioutil.ReadFile(logFile)
if err != nil {
return "", err
}
return string(buf), nil
}
// SendCoins attemps to send amt satoshis from the internal mining node to the
// targetted lightning node.
func (n *networkHarness) SendCoins(ctx context.Context, amt btcutil.Amount,
target *lightningNode) error {
balReq := &lnrpc.WalletBalanceRequest{}
initialBalance, err := target.WalletBalance(ctx, balReq)
if err != nil {
return err
}
// First, obtain an address from the target lightning node, preferring
// to receive a p2wkh address s.t the output can immediately be used as
// an input to a funding transaction.
addrReq := &lnrpc.NewAddressRequest{
Type: lnrpc.NewAddressRequest_WITNESS_PUBKEY_HASH,
}
resp, err := target.NewAddress(ctx, addrReq)
if err != nil {
return err
}
addr, err := btcutil.DecodeAddress(resp.Address, n.netParams)
if err != nil {
return err
}
addrScript, err := txscript.PayToAddrScript(addr)
if err != nil {
return err
}
// Generate a transaction which creates an output to the target
// pkScript of the desired amount.
output := &wire.TxOut{
PkScript: addrScript,
Value: int64(amt),
}
if _, err := n.Miner.CoinbaseSpend([]*wire.TxOut{output}); err != nil {
return err
}
// Finally, generate 6 new blocks to ensure the output gains a
// sufficient number of confirmations.
if _, err := n.Miner.Node.Generate(6); err != nil {
return err
}
// Pause until the nodes current wallet balances reflects the amount
// sent to it above.
// TODO(roasbeef): factor out into helper func
for {
select {
case <-time.Tick(time.Millisecond * 50):
currentBal, err := target.WalletBalance(ctx, balReq)
if err != nil {
return err
}
if currentBal.Balance == initialBalance.Balance+amt.ToBTC() {
return nil
}
case <-time.After(time.Second * 30):
return fmt.Errorf("balances not synced after deadline")
}
}
}